Coastal risk assessment of a micro-tidal littoral plain in response to sea level rise

Ocean & Coastal Management 104 (2015) 22e35

Contents lists available at ScienceDirect

Ocean & Coastal Management journal homepage: www.elsevier.com/locate/ocecoaman

Coastal risk assessment of a micro-tidal littoral plain in response to sea level rise Guido Benassai a, Gianluigi Di Paola b, *, Pietro Patrizio Ciro Aucelli c
Parthenope, Napoli, Italy Dipartimento di Ingegneria, Universita
del Molise, Pesche, IS, Italy Dipartimento di Bioscienze e Territorio, Universita c
Parthenope, Napoli, Italy Dipartimento di Scienze per l'Ambiente, Universita a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 April 2014 Received in revised form 20 November 2014 Accepted 20 November 2014 Available online

This paper illustrates an index-based coastal risk assessment that was performed on a micro-tidal alluvial plain taking into account the relative sea level rise (RSLR) for the evaluation of coastal vulnerability and exposure. This process took into account both the inundation of inshore land and the beach retreat due to storm surge, calculated on the basis of geomorphological data (bathymetry, sedimentology and beach width) and wave climate. The evaluation process was conceived with reference to a low and high hazard, associated with a wave storm with 1 year and 50 years return period. For the latter case, the response to RSLR was calculated taking into account both isostatic response and ice cap melting due to global warming, while the vertical land movement was assessed taking into account the different its rates in the northern and southern coastal area. The exposure and the damage of the coastal assets were evaluated with a simplified conceptual framework, which uses land cover data and a statistical population dataset. The risk assessment procedure was applied to Sele coastal plain, which involves numerous properties and important infrastructures, and is strongly susceptible to marine inundations. A sensitivity analysis of the vulnerability and the risk relative to different hazards and to RSLR was performed. Moreover, the final risk assessment classification was validated with a conceptual framework based on the observed damage ranking related to the tested coastal area. The obtained results showed that the northern high density urban areas were characterized by the highest risk, followed by some central areas with strong localized erosive focus. On the contrary, the southern zones, with wider beaches and almost intact dunes, were characterized by the lowest risk level. The results of this study were used for the development of a coastal protection project which, in fact, provided a different scheme for the coastline northwards and southwards of the Sele river mouth, according to the different risk ranking established. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Sele coastal plain Vulnerability assessment Risk assessment Relative sea level rise

1. Introduction As centers of human population, economic activities and social developments, coastal areas are frequently subject to natural hazards. Currently, 634 million people (10% of the world population) live in low-lying coastal regions 0e10 m above sea level. In addition, coastal populations are steadily increasing through migration (Mc Granaham et al., 2007). The Intergovernmental Panel on Climate Change (IPCC, 2014), argues that ongoing global change will cause the sea level to rise in coastal areas. Climate change

* Corresponding author. E-mail addresses: [email protected] (G. Benassai), gianluigi.dipaola@ unimol.it (G. Di Paola), [email protected] (P.P.C. Aucelli). http://dx.doi.org/10.1016/j.ocecoaman.2014.11.015 0964-5691/© 2014 Elsevier Ltd. All rights reserved.

generates sea level rise (SLR) for two main reasons, thermal expansion and ice cap melting. The most recent evidence suggests that global sea-level rise could reach 1 m or more during this century (Hansen and Sato, 2012; Grinsted et al., 2010; Pfeffer et al., 2008; Rahmstorf, 2007; Overpeck et al., 2006). These results are significantly beyond the upper limit of the range cited by the IPCC (2014): a 90% confidence interval between 26 and 82 cm. However, remote sensing analysis showed a lower increase rate for the Mediterranean basin in comparison to the global values, because the Mediterranean Sea is a semi-enclosed basin and therefore does not respond linearly to the influence of the open ocean (Vigo et al., 2011). The other climate-related effect in coastal zones besides SLR is the change in frequency, intensity and spatial pattern of coastal storms, which can have severe consequences for low-lying areas

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

prone to coastal and river flooding. Land subsidence and storm surges will result in permanent flooding of low-lying areas, inland extension of episodic flooding, increased beach erosion and saline intrusion (Cooper et al., 2008). Sea level rise, land subsidence and storm surges form an important disaster chain. Ongoing land subsidence amplifies sea level rise, which in turn tends to amplify storm surges and aggravates the disaster generated by storm surge (Karim and Mimura, 2008; Teatini et al., 2012). The inclusion of subsidence gives rise to the so-called Relative Sea Level Rise (RSLR), which is a key factor in risk management and planning at national, regional and local scales (Ramieri et al., 2011; Purvis et al., 2008; Bates et al., 2005). The term “vulnerability” is applied in different ways. In this paper we discuss the vulnerability of a specified system to a specified hazard or range of hazards (Adger et al., 2004). The term hazard refers specifically to a physical manifestation of climatic variability or change such as droughts, floods, storms, episodes of heavy rainfall, and so on. In this context the vulnerability is conceived in terms of the amount of damage caused to a system by a particular climate-related event or hazard (Jones and Boer, 2003). The hazards and impact approach typically views the vulnerability of a human system as determined by the nature of the physical hazard to which it is exposed, the frequency of occurrence of the hazard, and the system's sensitivity to the impacts of the hazard. This vulnerability, as a function of hazard and sensitivity, may be referred to as physical vulnerability (Brooks, 2003; Nicholls at al., 1999). Together with vulnerability and hazard, exposure is another pre-requisite of risk and disaster. Here, exposure is understood as the number of people and/or other elements at risk that can be affected by a particular event. In an uninhabited area the human exposure is zero, no matter the magnitude of the vulnerability and the hazard (Birkmann, 2007; Thywissen, K., 2006). While the vulnerability determines the severity of the impact an event will have on an element at risk, it is the exposure that drives the final tally of damage or harm, and so the final risk (IPCC, 2014). Among the different definitions of risk, in this paper we define the ‘risk’ as the probability of a loss, which depends on three elements, that is hazard, vulnerability and exposure. If any of these three elements in risk increases or decreases, then the risk increases or decreases respectively (Crichton, 1999). So a mitigation strategy for coastal risk can be achieved either by decreasing the vulnerability to flooding, or by decreasing the exposure from flooding or by combining the two (Luo et al., 2013; Pompe and Rinehart, 2008). According to the majority of the scientific community, coastal vulnerability is defined as the susceptibility of a socio-economic and/or ecological system to be harmed by a hazardous event (Cutter, 1996; Gornitz et al., 1994). Different semi-quantitative and quantitative procedures for its evaluation have been proposed in the literature. The first ones are mainly based on the subjective assessment of geomorphological indicators, while the second ones quantify the relative importance of physical and geomorphological relevant phenomena. The proposed methodologies have progressively evolved from single approaches, such as Bruun rule (Bruun, 1962) and UNEP methodology (Carter et al., 1994), to more recent consistent techniques, such as USGS-CVI (Gornitz et al., 1994) and SURVAS (Nicholls and de la Vega-Leinert, 2000), which provide an improved consideration of both physical and nonphysical factors. One of the most commonly used methods for assessing coastal vulnerability is based on the Coastal Vulnerability Index (CVI), which combines the changing susceptibility of the coastal system with its inherent response to a changing environment. The vulnerability classification is based upon the relative contributions and interactions of six variables, i.e., mean elevation, geology,

23

coastal landform, shoreline, wave height and tidal range (Thieler € et al., 1999; Gornitz et al., 1994, 1997). Ozyurt and Ergin (2010) developed a CVI to specifically assess the impacts induced by Sea Level Rise (SLR). The index is determined through the integration of five sub-indices, each one corresponding to a specific SLR related impact. With respect to the exposure evaluation, the socio-economic value of the elements at risk have been identified through the DPSIR framework (Driving forces-Pressure-State-ImpactResponses model), as recommended by the European Environmental Agency (EUROSION, 2004) and are reported in the literature Changes in the intensity and extent of economic activity determine the pressure trends and the multiple use demands on coastal resources. The “drivers” or “driving forces” lead to pressures on the environment, with, physical, biological and chemical repercussions. The changes in these conditions usually have environmental and economic impacts on ecosystems such as altered biodiversity or reduced resource availability, and ultimately on social and economic features of the society and human health as well. A set of appropriated societal and policy makers' prioritizations affecting any part of the chain between the drivers and the impacts can reduce undesired impacts (Rogers and Greenaway, 2005; Kristensen, 2004; Gabrielsen and Bosch, 2003). In this paper, starting from a recent approach for the Coastal Vulnerability Assessment (CVA) (Benassai et al., 2009, 2013; Di Paola et al., 2011), which is based on wave climate, bathymetry and sediment data, and evaluating the exposure on the basis of a simplified DPSIR scheme, we have assessed the risk of inundation and erosion of a test area for two different hazards: low hazard (1 year return period) and very high hazard (50 years return period) wave storm. For the latter case, the response to RSLR was evaluated taking into account the isostatic contribution, ice cap melting due to global warming and vertical land movement in the test area. The risk assessment procedure was applied to the Sele coastal plain, which is part of the coastal zone of Salerno province. It involves numerous properties and important infrastructures, and is very susceptible to marine inundations (Benassai et al., 2013; Vallefuoco et al., 2012). This coastal zone presents an intrinsic flooding vulnerability, because it is characterized by spread backridge depressions, with a mean height of about 0.50e1.50 m a.s.l. with a high sensitivity to future relative sea level rise (IPCC, 2014) and relative vertical land movements (Vilardo et al., 2009). In fact, Pappone et al. (2012), on the basis of stratigraphic and geochronological data, recognize different rates of vertical land movement for the northern and southern sectors of the Sele coastal plain. In particular, taking into account borehole stratigraphic data, the Authors evidence the following different behavior: the northern sector shows a mean rate of 0.4 mm/year from 5.0 ky BP to 1.0 ky BP, while the southern one is almost stable. Recently, Vilardo et al. (2009), analysing the satellite radar interferometry (PS-InSar) data, highlighted that this area is also affected by subsidence as several other Italian coastal plains (Cenni et al., 2013; Teatini et al., 2012; Baldi et al., 2009). The exposure of this area is also significant, due to concentration of an important inhabited area (city of Salerno), archaeological sites of UNESCO World Heritage List (ancient town of Poseidonia-Paestum) and important infrastructures located along the coastline. The paper is structured as follows: the morphological and climatological features, as well as the assessment of the RSLR of the study area are introduced in Section 2. The data and methods used to evaluate the coastal vulnerability, the exposure and the coastal risk are described in Section 3. Experimental results relevant to different hazards and the RSLR are presented and discussed in Section 4. Discussion and Conclusions are finally drawn in Section 5 and 6.

24

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

2. Study area 2.1. Beach morphology Sele plain is one of the widest alluvial coastal plains of Southern Italy, limited towards the sea by a narrow sandy beach which extends from NW to SE in the Gulf of Salerno, in the Southern Tyrrhenian Sea (Fig. 1A). This plain is the emerged continental portion of a large triangular-shaped morpho-tectonic depression, Salerno trough, related to the opening and expansion of the Tyrrhenian ocean basin started in the Upper Miocene (Amato et al., 2013; Casciello et al., 2006; Bartole et al., 1984). A younger coastal sector occurs between the Tyrrhenian sandy coastal ridge and the present shoreline. This belt represents the evolution of a barrierlagoon system, that shifted alternatively landwards and seawards during the Holocene. It includes a composite sand ridge system,

elevated 1e5 m a.s.l. which is partly exposed along the present coast and disappears inland under a muddy depression, rising about 1 m a.s.l. (Fig. 1B). During the last sixty years, the Sele coastline was affected by prevailing erosion that was very strong around the main river mouths, due to numerous hydraulic dams, which greatly reduced sediment supply to the rivers (Alberico at al. 2012a, 2012b; Pappone et al., 2011) as often is the case for many other coastal plains along the Tyrrhenian and Adriatic coasts (Rosskopf and Scorpio, 2013; Aucelli et al., 2009; Amorosi and Milli, 2001; Pranzini, 2001). The different morphological and anthropic features allow to distinguish between the more inhabited northern stretches of the coastline, which extend from the river mouth of Picentino to the river Asa, and the more intact coastal zones, reaching towards the South and the ancient town of Paestum. Here the area is characterized by wider beaches with almost perfectly preserved dunes.

Fig. 1. Study area (A) and topographical profiles with their location on coastal Sele Plain (B, C). The last one also shows the elevation areas under 2 m a.s.l. and under 1 m a.s.l., denoted in yellow and red, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

The beaches of the northern section are characterized by a strong anthropic impact, which is also confirmed by a number of sewage outlets protected by concrete structures, and some shore protection structures (small detached longitudinal and adherent breakwaters) placed here and there to protect single infrastructures and sometimes the coastal road. The beaches of the southern section, instead, are wider and show a lower anthropogenic load, with some exceptions on the left bank of river Sele, where a more intense impact is observed. At the southward limit of the physiographic unit, marked by the Capodifiume and Solofrone rivers, there are wider beaches with fine sand and a lower anthropic load. In order to define a detailed mapping of the emerged and submerged beach, 10 topographic and bathymetric profiles were produced (Fig. 1C). The topographic profiles were obtained by using a Differential Global Position System survey (DGPS), while the submerged beach profiles were obtained by means of a single-beam during a survey carried out by IAMC-CNR of Naples in 2008 and covered the area extending up to the closure depth (Di Paola et al., 2014).

2.2. Offshore and inshore wave climate Offshore wave climate was obtained through the statistical analysis of the data provided by the Italian Sea Wave Measurement Network (Rete Ondametrica Nazionale; RON, 2012) from the Ponza buoy (40
520 00.1000 N, 12
560 60.0000 E), which can be considered representative of the offshore wave conditions in the study area, according to the Wind and Wave Atlas of the Mediterranean Sea (Medatlas Group, 2004). The period covered by the wave data is July 1989eMarch 2008, including a total of 115,651 wave records, or sea states, each one characterized by a value of significant wave height, Hs, mean wave period, Tm, and mean wave direction, Dm. The results of the analysis show that the study area is frequently affected by moderate wave conditions associated to significant wave heights lower than 3 m, coming mainly from SSW-NNW sector. However, in the winter, stormy conditions are generated, mainly associated to wave fields traveling from subsector WSWWNW (Fig. 2). With regard to astronomical sea level variation, the study area experiences a typical semi-diurnal tide with a mean

25

tidal range of 0.45 m. However, main sea level variations due to meteorological surges can reach values up to 1 m (IIM, 2002). The long-term wave climate was extrapolated with the Weibull distribution. The equation for long term predictions of significant wave heights, with given return periods TR in years, is the following: 1

Hs;TR ¼ B þ A½lnðlTR Þ =k

(1)

where A, B and k represent the scale, position and shape factors respectively (Sarpkaya and Isaacson, 1981) and l is the mean number of “over-threshold” (Hs > 3 m) sea storms observed in one year, as available from statistical data (APAT, 2006). The spectral peak period Tp is determined by relevant statistical correlations between sea parameters. Table 1 lists the relevant parameters for TR ¼ 1 year and TR ¼ 50 years. The inshore wave climate used for risk calculations was obtained through the propagation of some significant wave storms on the beach profiles. These storms, recorded at Ponza buoy during winter 2010, are given in Table 2. A numerical wave model (SWAN) was used to simulate the wave storms and to propagate them on the beach profiles. It is a third-generation numerical wave model which describes temporal and spatial variation of wind-induced surface elevation, white-capping effects and friction with the sea bottom layer (Benassai, 2006). In SWAN, waves are described with the two-dimensional wave action density spectrum N]F/s, even when non-linear phenomena dominate (e.g., in the surf zone). The model is typically forced by using wind field forcing at 1 h intervals provided through the Advanced Research Weather Research and Forecast (WRF-ARW) wind field ECMWF model data. Outputs from SWAN model include significant wave height (Hs) on gridded fields, associated wave directions (Dm) and mean periods (Tm), as well as Table 1 Wave height prediction for storm surge with return period of 1 and 50 years. Parameter of Weibull distribution

TR ¼ 1 year

TR ¼ 50 years

A

B

k

l

Hs (m)

Tp (s)

Hs (m)

Tp (s)

0.89

3.51

1.00

5.87

5.1

9.8

8.6

12.2

Fig. 2. Wave rose (Hs-Dm) for the whole data set (a), for sea states with Hs > 4 m (b).

26

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

wave energy spectral information at different wavelengths (Benassai, 2006; Benassai and Ascione, 2006; Booij et al., 1999). The details of the simulations and the model validation with the buoy data are given in Benassai et al., 2013.

geomorphological analysis, based on cartographic (1:5.000 in scale) and photogrammetric interpretation. They identified a differential uplift/subsidence trend into the plain: an uplift of the whole Sele plain during the first part of the Holocene (up to ca. 6.0 ky BP) and a subsequent different behavior in the northern and southern sectors. The vertical land movements into the plain seem to continue until the present days, as evidenced by PS-InSar interpretation (1992e2001) of Vilardo et al. (2009). Two different values of vertical land movements were considered: 0.82 mm/year for the Dx Sele and þ0.2 mm/year for the Sx Sele plain, for the last 5 ky. The sum of the aforementioned parameters gave rise to the RSLR, reported in Table 3.

2.3. Sea level change assessment

3. Methodology of risk assessment

According to Lambeck et al. (2011), the three principal contributions to Relative Sea Level Rise (RSLR), measured with respect to land are (a) the sea level response to the past glacial cycle, including the response to the most recent glacial unloading of the major ice sheets, the response to the ocean floor loading by the melt water and the glacio-hydro-isostatic contributions, (b) changes in ocean volume in more recent times from thermal expansion and (c) vertical land movements (including uplift and subsidence). The isostatic response requires appropriate models for the past ice sheets, consisting in the inversion of rebound data from the formerly glaciated regions, and requires rheological parameters that are estimated from the sea level data for sites that are tectonically stable. Lambeck et al. (2011) calculated the value of 0.44 mm/year for the Sele alluvial plain (Table 3). The effects of the future change in oceans volume, caused by global warming and ice melting, has been evaluated by the last IPCC (2014) report, using the highest and worst sea level rise predictions. We have adopted the maximum and minimum projections, known as RCP2.6 and RCP8.5, evaluated for the increase of global sea mean surface temperature of 1
C and 2
C, respectively. For each of the scenarios we considered the mean values of 4.6 mm/year and 5.8 mm/year (Table 3). These results are in line with those of the analysis of the Topex-Poseidon satellite altimetry data in the period 1993e1998 (Cazenave et al., 2001), which give an average rate of 5e10 mm/year in the western Mediterranean basin. However, more recent additional remote sensing data for 1992e2008 period (Vigo et al., 2011), showed an overall increase of 2.6 cm, which should give a lower increase rate of approximately 2 mm/year for the same basin. The Authors acknowledge that this change is not consistent with the accepted global rate of 3.0e3.5 mm/year since 1992 (Ablain et al., 2009; Cazenave et al., 2009; Beckley et al., 2007), but they argue that the Mediterranean Sea is a semi-enclosed basin which does not respond linearly to the influence of the open ocean. The vertical land component of the Sele alluvial plain was evaluated by Pappone et al. (2012). The Authors considered a chrono-stratigraphic study, integrated by a detailed

This Section presents the proposed two-step procedure used to evaluate the risk due to coastal erosion and inundation along the Sele coastal plain, taking into account the RSLR. First, we applied the Coastal Vulnerability Assessment (CVA), then we evaluated the exposure and consequently the coastal risk of the test area for two different hazards: low hazard (1 year return period) and very high hazard (50 years return period) wave storm. For the latter case, the response to RSLR was evaluated in the test area with the procedure described in section 2.3. The first step (sub-section 3.1) involves the evaluation of the coastal vulnerability on the considered profiles. The second step (sub-section 3.2) regards the evaluation of the exposure, calculated through the socio-economic value and the damage value. Finally, in the sub-section 3.3, the risk assessment is calculated according to the European Union Commission (ISO/IEC, 2009).

Table 2 Recorded wave storms of winter 2010. Storm n.

Duration

Hs max (m)

Tp max (s)

Dm max (
N)

TD (h)

1 2 3

09/11/10e10/11/10 17/12/10e18/12/10 23/12/10e25/12/10

4.23 5.01 4.29

9.5 9.5 10.0

218 231 255

55 24 48

Table 3 Sea level rise scenarios according to IPCC (2014), using two different projections for the 2065. The values were calculated for the next 51 years (2014e2065). Scenarios (2065)

RCP2.6 RCP8.5

Sectors of Sele plain

Dx Sele Sx Sele Dx Sele Sx Sele

Change in ocean volume

Glaciohydroisostasy

Land movement

RSLR

mm/ year

mm mm/ year

mm mm/ year

mm

Mm

0.44 0.44 0.44 0.44

21.9 21.9 21.9 21.9

240 0.82 240 0.2 300 0.82 300 0.2

41.8 10.2 41.8 10.2

303.7 251.7 363.7 311.7

4.6 4.6 5.8 5.8

3.1. Coastal vulnerability assessment The coastal vulnerability was evaluated, starting from the CVA method already proposed by Di Paola et al. (2014) and Benassai et al. (2013), with the inclusion of RSLR in order to examine its influence on the beach response. We adopted the RCP8.5 scenario relative to 2065, which was added to the very high hazard of TR ¼ 50 years, summing up the different components described in section 2.3. The CVA methodology starts with the identification of key variables representing significant driving processes, then it attributes different semi-quantitative scores according to a 1e4 scale; 1 indicates a low contribution to coastal vulnerability of a specific key variable for the study area or sub-area, while 4 indicates a high contribution. Then the key variables are integrated in a single index according to the following equation:

CVA ¼ IRu þ IR þ ID þ E þ T

(2)

where IRu is an index related to wave run-up distance on the shoreline, IR is an index linked to the short-term erosion, ID is the coastal structures stability index, E is the beach erosion rate index and T is the tidal beach response. The Coastal Vulnerability Assessment was carried out by evaluating equation (2) without considering ID and T contributions. In fact, the test area (i.e., the southern Tyrrhenian Coastal Sea basin) is a micro-tidal coastal environment with a max tidal excursion of 0.45 m (IIM, 2002) (hence, T ¼ 0), where no coastal protection is present (hence, ID ¼ 0). Therefore, only IRu, IR, and E contributions have been taken into account for the evaluation of Coastal Vulnerability Assessment. For each index, the relative score was assigned to ranks 1, 2, 3 and 4 from very low to high. The resulting index CVA was obtained

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

by the simple addition of the single indices, according to EUROSION project (EUROSION, 2004). The Wave run-up height index (IRu) provides a measurement of the potential inundation capacity, which characterizes natural beaches with respect to wave storms, through the evaluation of the horizontal distance traveled by the wave in the run-up process (Xmax), that is the run-up level divided by foreshore beach slope (Stockdon et al., 2006). IRu values for inundation rate of the natural beach were customarily clustered into four discrete levels i.e. stable (IRu ¼ 1), low (IRu ¼ 2), moderate (IRu ¼ 3) and high (IRu ¼ 4) (Table 4). The Short-term erosion index (IR) provides a measurement of the maximum beach retreat (Rmax) calculated with Kriebel and Dean (1993) convolution method and normalized with the beach with L. IR values for short-term erosion of natural beach were customarily clustered into four discrete levels i.e., stable (IR ¼ 1), low (IR ¼ 2), moderate (IR ¼ 3) and high (IR ¼ 4) (Table 4). In order to compute the beach erosion rate index (E), which provides the evaluation of historical shoreline change, a number of topographic maps, aerial photographs and multi-spectral satellite images were used to demarcate shoreline positions in different periods. The image distortion in the aerial photographs due to tilt and pitch of aircraft, and the scale variations caused by changes in altitude along the flight lines were corrected by Z-Map (MENCI™) software. The error related to orthorectification was controlled through the root mean square error (RMSE), which was ±3 m. The historical maps were georeferenced using a number of Ground Control Points (GCPs), which were extracted by comparing each historical map with the 1998 georeferenced Campania technical map with 1:5000 scale. Moreover, because it was not possible to reconstruct tidal conditions at the time of the photo shot, it was assumed that the daily water line position was subject to a maximum uncertainty of ±1.6 m, taking into account the daily mean sea-level change of ±20 cm (ISPRA, 2013) and the mean intertidal slope of 12% calculated on the beach profiles. The beach erosion rate index was customarily clustered into discrete levels depending on the erosion rate velocity E (m yr1), i.e., stable (E ¼ 1), low (E ¼ 2), moderate (E ¼ 3) and high (E ¼ 4) (Table 4). In order to take into account the RSLR for the vulnerability calculation, we used the RCP8.5 projection in the Davidson-Arnott (2005) flooding method to estimate the potential coastal retreat for 2065. This model assumes that the shoreline displacement is related to the beach slope (tan b) as follows:

1 Rb ¼ S$ tan b

(3)

where Rb is the shoreline retreat, S is the RSLR and b the slope of the profile between the dune ridge and the closure depth. So, for each profile a beach response index due to RSLR (IRSLR) was introduced as:

Table 4 Classification of IR, IRu, E, IRSLS and CVA ranking. Variables

IR (%) IRu (%) E (m/year) CVA (%) IRSLR (%) CVARSLR (%)

Stability

Low

Moderate

High

1

2

3

4

15 40 0.5 ≤30 25 ≤30

16e30 41e60 0.6e1.0 31e50 26e50 31e50

31e50 61e80 1.1e2.0 51e70 51e75 51e70

>50 >80 >2.0 >70 >75 >70

IRSLR ¼

Rb $100 L

27

(4)

where Rb is shoreline retreat trough Davidson-Arnott (2005) formula and L is the beach width. Consistently, we have introduced a new Coastal Vulnerability Assessment index, given by:

CVARSLR ¼ CVA þ IRSLR

(5)

Also IRSLR values were customarily clustered into four discrete levels i.e., stable (IRSLR ¼ 1), low (IRSLR ¼ 2), moderate (IRSLR ¼ 3) and high (IRSLR ¼ 4) (Table 4). The CVARSLR index was evaluated according to equation (5). The CVA and CVARSLR consequently obtained were clustered into the four categories already introduced (Table 4). 3.2. Coastal exposure The method proposed in this paper aims at representing the socio-economic pressure factors as well as the exposed value of the assets located in the coastal area. It is based on a simplified Driving forces e Pressure e State e Impact e Responses model (DPSIR) as recommended by the European Environmental Agency and already used by EUROSION (2004). The concept of Radius of Influence of Coastal Erosion (RICE) is introduced as the total area within 500 m from the shoreline and lying below 5 m above sea level. Once the RICE has been defined, the approach considers a number of indicators, which are based on the population density and on the coastal land use. In the following, the socio-economic value (percentage of loss of an element or a group of elements that should occur in case of hazard) and the exposed value (quantification of the receptors that can be influenced by the hazard) are developed in detail. 3.2.1. Socio-economic value Based on EUROSION (2004) and FLOODsite (2004), the socioeconomic value (S) was evaluated using the coastal socioeconomic value index (CSI), which depends on the land use characteristics and the total population of the municipalities included in the test area. The land use data were obtained by maps realized by Campania Region in 2009 (CUAS, 2009) while the statistical data were obtained through the last ISTAT dataset (ISTAT, 2011). Moreover, the ISTAT (2001) dataset was also used as a comparison with ISTAT (2011) dataset in order to evaluate the increase of inhabited areas. CSI is given by the following equation:

CSI ¼ PRICE þ URICE þ ERICE þ U10km

(6)

where PRICE is the resident population in the RICE area, URICE is the percentage of urbanized/industrialized area in RICE, ERICE the percentage of ecological value area in RICE, while U10km is the percentage of the increased urbanized area inside RICE. The resident population in RICE (PRICE) was calculated from the formula:

PRICE ¼ DU $AURICE þ DR $ARRICE þ DN $ANRICE

(7)

where AURICE, ARRICE and ANRICE denote the urban, rural and natural area in RICE, DU, DR and DN are the demographic density for urbanized, rural and natural areas. The urban, rural and natural areas were evaluated by the land use classification reported in CUAS (2009), while the demographic densities for each area were obtained by EUROSION (2004) estimates, further modified by ISPRA

28

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

(2009). In particular, these densities were obtained as a portion to the total municipality population, according to the following equations:

DU ¼ 0:7485 Pop=AU ;

DR ¼ 0:217 Pop=AR ;

DN

Table 6 Ranking of socio-economic value (S) used to evaluate CSI. CSI 2 S1

S

3e5 S2

6e8 S3

9 S4

¼ 0:035 Pop=AN ; in which AU, AR and AN represent the urbanized, rural and natural areas, while Pop is the municipality population, extracted by the ISTAT (2011) dataset. The urbanized/industrialized index (URICE) and the ecological value index (ERICE) are the urban and ecological areas in RICE in percentage to the total RICE area, while the U10km index is the incremental percentage of urban/industrial area within 10 km of the coastline during the decade 2001e2011 (ISTAT, 2011, 2001). Table 5 shows the scores assumed for each parameter, which were customized in a classification slightly different from EUROSION (2004), as done by Martinelli et al. (2010). Based on the scores reached by PRICE, URICE, ERICE, U10km, and according to eq. (6), the CSI values were obtained, so the corresponding socio-economic values (S) were evaluated and reported in Table 6. In detail S4 denotes a very high socio-economic value, S3 a high socio-economic value, S2 a medium socio-economic value and S1 a low socio-economic value. 3.2.2. Potential damage levels The potential damage of the assets located in the coastal area depends on the type of asset, and on the method used to determine the asset value. Some examples of valuation methods include the present value, replacement cost, damages avoided and output value. For example, the value of a beachside park can be related to the replacement costs; likewise, there are a range of other values in addition to the monetary value including the social, health and lifestyle values offered through the provision of a coastal infrastructure. In this view, the potential damage (D) represents the social, economic and natural value of the assets located in the coastal area in terms of monetary costs or human lives. In order to evaluate these costs it is necessary to take into account the monetary, environmental, cultural and archaeological values in the study area. In order to classify the assets located in the coastal area, the latter was divided into in 13 zones (Z), according to the municipality borders and to buildings, infrastructures and natural areas extracted from the 2004 orthophoto and from the 1998 Campania technical map. The following damage levels have been considered: D4 (very high damage) for groups of inhabited houses with roads and permanent infrastructures (urbanized areas); D3 (high damage) for localized inhabited houses, special zones of conservation and protection, national heritage protection areas and public interest buildings;

Table 5 Classification and ranking of PRICE, URICE, ERICE and U10km used for CSI evaluation. Indicator

0

1

30

e

10e20% >20%

e

D2 (medium damage) for spread houses, rural areas, touristic facilities and dune ridge; D1 (low damage) for natural (unprotected) areas.

3.2.3. Exposure levels The coastal exposure (Ex) denotes the percentage of the loss of an element or a group of elements that could occur in case of hazard. This parameter is evaluated by the matrix shown in Table 7. In particular the following scores were considered: Ex4 (very high exposure): potential human loss and high economic loss or damage; Ex3 (high exposure): possible human losses and economic issues; Ex2 (medium exposure): minor human loss threats and lower economic losses; Ex1 (low exposure): no human loss danger and/or negligible damage. These different levels have been obtained as a matrix product of socio-economic value by the damage level as shown in Table 7. 3.3. Coastal risk levels The European Union Commission (ISO/EC, 2009) defines the risk as “the probability of harmful consequences, or expected losses (deaths, injuries to property, livelihoods, disruption to economic activities or environment), resulting from interaction between vulnerability and exposure”. Therefore, the risk can be obtained through the following definition:

R ¼ V$Ex ¼ V$ðS$DÞ

(8)

where V is the coastal vulnerability (susceptibility of a coastal area to be affected by either inundation and/or erosion), Ex is the exposure described in section 2.1, S is the socio-economic value (percentage of loss of an element or a group of elements that should occur in case of hazard), and D is the potential damage (concerning the monetary, human life and environmental values in the coastal area). The risk level (R) was evaluated by eq. (7), given by the matrix product indicated in Table 8, which determines the risk level through the product of the coastal vulnerability by the exposure. In particular, the risk level has been classified as follows: R4 (very high risk): high danger for human life and/or permanent loss of structures or social and financial activities; Table 7 Coastal exposure (Ex) evaluation, given by the product matrix of socio economic value index (S) by potential damage level (D).

D4 D3 D2 D1

S4

S3

S2

S1

Ex4 Ex3 Ex2 Ex1

Ex3 Ex2 Ex1 Ex1

Ex2 Ex1 Ex1 Ex1

E x1 E x1 E x1 E x1

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

29

Table 8 Risk level (R) evaluation, by matrix product of coastal vulnerability (V) by exposure (Ex).

R3 (high risk): significant damage to the population and/or loss of structures and buildings; R2 (medium risk): only loss to the residential properties and to the infrastructures; R1 (low risk): possible, but low social, financial and economic damages.

4. Experimental results 4.1. Coastal vulnerability evaluation

Fig. 3. Partial contributions of IR, IRu and E to the CVA index for TR ¼ 1year (A) and TR ¼ 50 years (B).

The Coastal Vulnerability Assessment has been calculated with reference to a low and high level hazard, that is, a storm with 1 year and 50 years return period, in the latter case with the additional beach response to the RSLR. The following results can be highlighted.  CVALH, refers to low hazard (wave storm with TR ¼ 1 year) and shows that the beach profile P9 is the only one with a low vulnerability level (V1). The beach profiles P8eP10, P1eP2eP5 and P3eP4eP6eP7 have a medium (V2), high (V3), and very high (V4) vulnerability level, respectively (Fig. 3A).  CVAHH, refers to high hazard (wave storm with TR ¼ 50 years), and shows that the beach profile P9 changes its vulnerability level from low (V1) to medium (V2). The profiles P8 and P10 change their vulnerability from medium (V2) to high (V3), the profiles P2 and P5 from high (V3) to very high (V4), while the profiles P1, P3, P4, P6 and P7 leave their vulnerability levels unchanged. (Fig. 3B).  CVARSLR, refers to high hazard with the inclusion of RSLR, and was calculated through the estimate of the beach retreat (Rb) associated to RCP8.5-2065 projections (see section 2.3). This parameter is reported in Fig. 4A, while the non-dimensional values are given in Fig. 4B. The examination of Fig. 4 shows that several transects are exposed to RSLR, particularly the P7 profile, where the beach retreat for the scenario RPC8.5 is greater than the present beach width, while a lower impact is noted for the southern part of the test area (P8, P9 and P10 profiles). The comparison between Coastal Vulnerability Assessment relative to different hazard levels shows that the rise in the hazard level increases the vulnerability score of only half of the profiles, because the central profiles have already reached a very high vulnerability level (V4). The comparison between CVAHH (Fig. 4B) and CVARSLR (Fig. 5) shows that RSLR raises the vulnerability level from V3 to V4 only for profile P1, while the rest of the profiles leave their vulnerability unchanged.

Fig. 4. Calculated value of beach retreat (Rb) associated with RCP8.5 projections in comparison with the present beach width (A); corresponding values of IRSLR (B).

30

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

4.2. Exposure evaluation

Fig. 5. Calculated values of CVA, with the comparison between IRSLR and the other vulnerability parameters.

According to Table 5, the exposure (Ex) is given as the socioeconomic value (S) multiplied by the potential damage (D). The socio-economic value (S) was evaluated on the basis of the CSI values as reported in Table 6. Based on the statistical information of the socio-economic parameters for CSI, and on the land use CUAS (2009) classification, a schematic representation of main land uses was obtained and given for each municipality in the pie charts of Fig. 6. In Table 9 the calculations of PRICE, URICE, ERICE and U10km were reported in detail and classified according to the ranking given in Table 5. The highest values PRICE and URICE were obtained for the Salerno municipality because it is the most inhabited town in the Sele plain (over 20,000 people), which exhibits a very high urbanized percentage URICE (91.5%). Moreover, the higher ERICE values for Eboli and Capaccio were obtained because they are included in protected areas. Finally, U10km resulted to be higher for Salerno and Eboli, very low for Battipaglia and almost absent for Capaccio.

Fig. 6. Coastal potential damage map with land use results for each municipality on the Sele plain.

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

31

Table 9 Evaluation of PRICE, URICE, ERICE and U10km for each municipality of Sele plain. Moreover, the resulting CSI index and socio-economic value S are reported. Municipality

Salerno Pontecagnano Battipaglia Eboli Capaccio

URICE

PRICE

ERICE

U10km

Pop

Index

%

Index

%

Index

%

Index

CSI

S

26,425 3555 2313 1969 3362

4 3 2 2 3

91.5 30.4 30.2 16.8 31.0

4 2 2 1 2

0.0 0.0 0.0 72.0 44.3

0 0 0 3 3

21.7 17.7 5.2 26.7 1.2

3 2 1 3 0

11 7 5 9 8

S4 S3 S2 S3 S3

The potential damage (D) was evaluated based on the subsection 3.2.2, in which the highest damage is for inhabited and protected areas, the lowest for uncultivated areas. This is the reason why the highest value of D was obtained for Z1, Z2, Z11 and Z12 zones, of which the first two are associated to a very crowded area, while the other two are associated to archaeological and protected zones. The rest of the areas have a medium damage, while Z5 and Z6 zones, associated to an uncultivated area, have a low damage level (Fig. 6). The exposure (Ex) has been obtained with the matrix product of S multiplied by D, therefore the highest Ex values are associated to Z1, Z2, Z3 and Z4 zones, which are characterized by simultaneous high values of S and D (for anthropic and/or environmental reasons). The central part of the coastal area, corresponding to Z3, Z7, Z8, Z9, Z10 and to the last area Z13, is associated to a medium level of exposure Ex2. Finally, the rural and less inhabited coastal stretches corresponding to profiles Z4, Z5 and Z6 are associated to a low exposure Ex1 (Fig. 6).

4.4. Risk validation through damage detection The obtained risk levels were validated through a visual damage assessment for each coastal zone included in the considered area. The flood damage on each part of the coastal stretch was detected on the basis of information taken from different sources, like newspapers and TV reports, validated with field surveys. This is a common practice when localized damages are concerned, otherwise a comparison between remote sensing images before and after the flooding event is needed. In this paper we classified the Observed Damage (OD) on the basis of the structures involved. In particular, we established the following four rankings: OD4 (very high observed damage): buildings and infrastructures; OD3 (high observed damage): touristic facilities, dune ridges, cultural heritage; OD2 (medium observed damage): natural zones, cultivated areas; OD1 (low observed damage): not significantly damaged areas.

4.3. Risk evaluation Based on the assumed ranking for the different scores of coastal vulnerability (V) obtained by the CVA method (Figs. 4A and 6), and based on the exposure (Ex) ranking given in Fig. 6, we obtained the coastal risk results given in Fig. 7A and B, for a low (TR ¼ 1 year) and high (TR ¼ 50 years) hazard coupled also with RSLR related to 2065. Being the risk given by the matrix product of the coastal vulnerability for the exposure, it assumes the highest values only when both factors are highest. This is not the case for the low hazard (storm associated with TR ¼ 1 year), because the highest vulnerability values V4 are coupled with low Ex1 and Ex2 exposure values (Z4, Z5, Z6, Z8, Z9 and Z10 zones), while the highest exposure Ex4 is coupled with V3 value only in Z1 zone. However, for the high hazard with TR ¼ 50 years, the maximum risk level R4 is reached (in Z1 zone, followed by R3 in Z2 zone), because the highest Ex4 exposure value is associated with V4 and V3 vulnerability levels. The examination of Fig. 7A shows that the risk is maximum for Salerno municipality (Z1), lower in Pontecagnano, Battipaglia and Eboli (from Z2 to Z7), then it slightly increases closer to the boundary of Capaccio municipality (Sele mouth e Z8) and decreases again for the rest of the coastal areas. The comparison between Fig. 7A and B shows that the increase in hazard in most cases raises the risk level by one score (Z1, Z2, Z3, Z8, Z11, Z12 and Z13) except in Z4, Z5 and Z6 zones where the vulnerability is already at a maximum (V4). In particular, the low risk ranking relevant to Z3 zone increases to medium due to the increase of the vulnerability, which rises from V3 to V4. The low risk ranking relevant to Z7 and Z11 zones increases to medium due to the increase in vulnerability levels, which rises from V3 to V4. The low risk level is unchanged in Z4, Z5, Z6, Z12 and Z13, because the limited or absent vulnerability increase is coupled with a low or medium exposure.

The different observed damages are reported in the photographs of Fig. 8, together with their ranking. In particular, the observed damage in Z1 and Z2 zones (Fig. 8A and B) is relevant to building and infrastructures (OD4), the observed damage in Z4 and Z5 zones is absent (Fig. 8C), while the zones Z8 and Z9 (Fig. 8D and E) are characterized by damages to the dune ridges and to some touristic facilities (OD3). Finally, the zones Z12 and Z13 are characterized by an almost intact dune ridge with no significant damage (Fig. 8F and G). The comparison between the risk level and the observed damage for each zone shows that the highest risk values (R3 and R4) are given only for Z1 and Z2 zones, where a maximum exposure value (due to highly inhabited area) is coupled with a very high vulnerability V4. This high risk level is validated by a high observed damage level OD4. In the central stretches of the coastline, the risk level decreases towards a medium-low level, which is validated by few observed damages. The damages were experienced instead in Z8 and Z9 zones (Sele mouth), where the higher risk is validated through a high observed damage level ranking OD3. In the southern part of the coastline, the lower risk ranking is validated by a low observed damage level OD1.

5. Discussion The present Coastal Risk Assessment method (CRA) was based on the previous Coastal Vulnerability Assessment Method (CVA) through the evaluation of the exposure, given by the product of the socio-economic value S and the potential damage value D. In order to lower the subjectivity of the method, we avoided to introduce indices based upon a weighted average, because the weights reflect the importance of each factor to the final decision, which often requires subjective assessment. Rather we chose to

32

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

Fig. 7. Coastal risk map of Sele plain with reference to TR ¼ 1 year (A) and TR ¼ 50 years (B).

obtain a vulnerability and risk index that summed up the scores of each index, thus leaving the subjectivity only to the mechanism of the matrix product, which is straightforward. In order to examine the strengths and weaknesses of the proposed methodology, a sensitivity analysis of the Coastal Vulnerability Assessment to the examined hazard level and to the RSLR projections was performed, together with the CRA validation in most zones based on the comparison between the risk level and the observed damage ranking. The Coastal Vulnerability Assessment sensitivity to the hazard level is given by the comparison between the results relative to a low (TR ¼ 1 year) and high (TR ¼ 50 years) hazard level (Fig. 7A and B). It shows that half of the profiles increase their vulnerability score, while the rest of the profiles (mainly the ones of the central part of the coastal stretch) do not exhibit an increase. The Coastal Vulnerability Assessment sensitivity to RSLR is lower than that to the hazard level, in fact the inclusion of RSLR in the high hazard raises the vulnerability level from V3 to V4 only in profile P1. The increase in RSLR projections would certainly increase the CVAHHRSLR for P9 profile, while the V3 score for profiles P8 and P10 would remain unchanged. The Coastal Risk Assessment sensitivity to RSLR is given as a consequence of the Coastal Vulnerability Assessment analysis. The

increase in vulnerability due to RSLR has a corresponding increase in risk level in the northern profile P1, where the sea level rise effect is amplified by subsidence. When the RSLR is further increased, the risk level for P9 profile would further increase from R1 to R2. As a matter of fact, the rest of the profiles should maintain their vulnerability levels, being already maximum, and therefore should also maintain their risk level. The results of the coastal risk assessment need a proper calibration, otherwise significant differences may occur between the risk levels and the real damages which occurred on the different coastline stretches. The lack of calibration may lead also to significant differences between the results of similar approaches on the same coastal stretch. As an example, Martinelli et al. (2010) obtained coastal risk maps for Emilia Romagna, using the same slightly modified EUrosion approach and another similar approach based on multiple regression analysis (EUROSION, 2004). They obtained the risk ranking in the same way, summing up the different indices and assuming four risk levels (low, moderate, high and very high) according to the categories: R1 < 10, 10 < R2 < 20, 20 < R3 < 30, R4 > 30. The results obtained were compared with the ones achieved with the multiple regression analysis, with significant differences in the low and high risk areas. These discrepancies were explained with the need of a proper calibration.

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

33

Fig. 8. Coastal observed damages classification in comparison of the coastal risk map (TR ¼ 50 years).

In order to calibrate our risk level results, we compared the different risk ranking with the observed damage ranking in most coastal zones. The risk ranking on Z1 and Z2 zones (Salerno and Pontecagnano municipalities) was validated by observed damages to buildings and to the coastal road (OD4), while the Z3, Z4, Z5 and Z6 zones were not significantly affected by coastal damages. The calculated risk level of the Z8 and Z9 zones was validated through the observed damage to touristic facilities of a holiday beach resort and to the dune ridge (OD3). For the southern zones, no significant

permanent damage was observed, thus validating a lower risk score (OD1). The project of the defense system for the Sele coastal plain, implemented by the Province of Salerno for a total cost of more than 40 million Euro, is in line with the present risk evaluation results. As a matter of facts, the project philosophy is different for the northern and southern part of the littoral stretch. In the northern littoral portion, which coincides with the municipality of Pontecagnano, extending for a total length of about 4.0 km, the project consisted in the realization of a cell defense

34

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35

system, made of partially submerged groins connected offshore by submerged longitudinal breakwaters for the entire coastal length. In the stretch of coastline located southern of the Sele river mouth the coastal defense works have been reduced as much as possible due to the relative stability of this stretch of shoreline. The littoral, however, is of high environmental and archaeological value, so the project consisted in a vast monitoring program with some localized nourishment interventions, particularly at the Sele river mouth, which is in localized erosion due to scarcity of river sediments. 6. Conclusions In this paper, an index-based coastal risk assessment was performed on a micro-tidal alluvial plain taking into account the RSLR. The Coastal Risk Assessment (CRA) was evaluated on the basis of the vulnerability calculation with the Coastal Vulnerability Assessment (CVA) method and on the coastal potential damage evaluation through land use maps and statistical information on the population density. The CVA sensitivity analysis shows that the results are more influenced by the hazard level than the RSLR inclusion. In fact, the RSLR increases the coastal vulnerability mainly for the northern profiles, where the sea level rise effect is amplified by the subsidence, while the hazard level increases the vulnerability of both the northern and central southern profiles. The validation of the CRA method is satisfactory because most of the zones classified with a moderate/high risk are also classified with a high/very high observed damage ranking. The northern high density urban areas are characterized by the highest risk, followed by some central areas with strong localized erosive focus. On the contrary, the southern zones, with wider beaches and almost intact dunes, are characterized by the lowest risk level. These results suggested the development of a defense system for the entire coastal plain, with different levels of interventions according to the risk levels. The northern coastal stretch is going to be protected with a series of submerged T-shaped and longitudinal breakwaters, while the southern coastal stretch is going to be interested by nourishment interventions, to be gradually realized with a monitoring program. References Ablain, M., Cazenave, A., Guinehut, S., Valladeau, G., 2009. A new assessment of global mean sea level from altimeters highlights a reduction of global slope from 2005 to 2008 in agreement with in-situ measurements. Ocean Sci. 5, 193e201. Adger, W.N., Brooks, N., Kelly, M., Bentham, S., Eriksen, S., 2004. New Indicators of Vulnerability and Adaptive Capacity. Tyndall Centre Technical Report 7. Alberico, I., Amato, V., Aucelli, P.P.C., D'Argenio, B., Di Paola, G., Pappone, G., 2012a. Historical shoreline evolution and recent shoreline trends of Sele Plain coastline (southern Italy). The 1870e2009 time window. J. Coast. Res. 28, 1638e1647. http://dx.doi.org/10.2112/JCOASTRES-D-10-00197.1. Alberico, I., Amato, V., Aucelli, P.P.C., Di Paola, G., Pappone, G., Rosskopf, C.M., 2012b. Historical and recent changes of the Sele River coastal plain (Southern Italy): natural variations and human pressures. Rend. Lincei-Sci. Fis. 23, 3e12. http:// dx.doi.org/10.1007/s12210-011-0156-y. Amato, V., Aucelli, P.P.C., Ciampo, G., Cinque, A., Di Donato, V., Pappone, G., Petrosino, P., Romano, P., Rosskopf, C.M., Russo Ermolli, E., 2013. Relative sea level changes and paleogeographical evolution of the southern Sele plain (Italy) during the Holocene. Quatern. Int. 288, 112e128. http://dx.doi.org/10.1016/ j.quaint.2012.02.003. Amorosi, A., Milli, S., 2001. Late Quaternary depositional architecture of Po and Tevere river deltas (Italy) and worldwide comparison with coeval deltaic successions. Sediment. Geol. 144, 357e375. APAT, 2006. Atlante Delle Onde Nei Mari Italiani. Universit
a degli Studi di Roma Tre, Roma. Aucelli, P.P.C., Iannantuono, E., Rosskopf, C.M., 2009. Evoluzione recente e rischio di erosione della costa molisana (Italia meridionale). Ital. J. Geosci. (B. Soc. Geol. Ital.) 128, 759e771.

Baldi, P., Casula, G., Cenni, N., Loddo, F., Pesci, A., 2009. GPS-based monitoring of land subsidence in the Po Plain (Northern Italy). Earth. Planet. S. C. Lett. 288, 204e212. http://dx.doi.org/10.1016/j.epsl.2009.09.023. Bartole, R., Savelli, D., Tramontana, M., Wezel, F.C., 1984. Structural and sedimentary features in the Tyrrhenian margin off Campania, Southern Italy. Mar. Geol. 55, 163e180. Bates, P.D., Dawson, R.J., Hall, J.W., Horritt, M.S., Nicholls, R.J., Wicks, J., Hassan, M.A.M., 2005. Simplified two-dimensional numerical modeling of coastal flooding and example applications. Coast. Eng. 52, 793e810. Beckley, B.D., Lemoine, F.G., Luthcke, S.B., Ray, R.D., Zelensky, N.P., 2007. A reassessment of global rise and regional mean sea level trends from TOPEX and Jason-1 altimetry based on revised reference frame and orbits. Geophys. Res. Lett. 34, L14608. http://dx.doi.org/10.1029/2007GL030002. Benassai, G., 2006. Introduction to Coastal Dynamics and Shoreline Protection, first ed. Wit Press, Southampton. Benassai, G., Ascione, I., 2006. Implementation and validation of wave watch III model offshore the coastlines of Southern Italy. In: OMAE, vol. 2006e92555, pp. 1e8. Benassai, G., Chirico, F., Corsini, S., 2009. Una metodologia per la definizione del rischio da inondazione costiera. Studi Costieri 16, 51e72. Benassai, G., Montuori, A., Migliaccio, M., Nunziata, F., 2013. Sea wave modeling with X-band COSMO-SkyMed© SAR-derived wind field forcing. Ocean. Sci. 9, 325e341. http://dx.doi.org/10.5194/os-9-325-2013. Birkmann, J., 2007. Measuring Vulnerability to Natural Hazards: towards Disaster Resilient Societies. United Nations University, Tokyo. Booij, N., Ris, R.C., Holthuijsen, L.H., 1999. A third-generation wave model for coastal regions 1: model description and validation. J. Geophys. Res. 104, 7649e7666. http://dx.doi.org/10.1029/98JC02622. Brooks, N., 2003. Vulnerability, Risk and Adaptation: a Conceptual Framework. Tyndall Centre Working Paper No. 38. Bruun, P., 1962. Sea-level rise as a cause of shore erosion. J. Waterw. Harbours Div. 88, 117e130. Carter, T.R., Parry, M.L., Nishioka, S., Harasawa, H., 1994. Technical Guidelines for Assessing Climate Change Impacts and Adaptation. University College London, London, and Centre for Global Environmental Research, Tskuba. Casciello, E., Cesarano, M., Pappone, G., 2006. Extensional detachment faulting on the Tyrrhenian margin of the southern Apennines contractional belt (Italy). J. Geol. Soc. London 163, 617e629. Cazenave, A., Cabanes, C., Dominh, K., Mangiarotti, S., 2001. Recent sea level change in the Mediterranean Sea revealed by satellite altimetry. Geophys. Res. Lett. 28, 1607e1610. Cazenave, A., Dominha, K., Guinehutb, S., Berthiera, E., Llovela, W., Ramilliena, G., Ablainb, M., Larnicol, G., 2009. Sea level budget over 2003e2008: a revaluation from GRACE space gravimetry, satellite altimetry and Argo. Glob. Planet. Change 65, 83e88. http://dx.doi.org/10.1016/j.gloplacha.2008.10.004. Cenni, N., Viti, M., Baldia, P., Mantovanib, E., Bacchettia, M., Vannucchi, A., 2013. Present vertical movements in Central and Northern Italy from GPS data: possible role of natural and anthropogenic causes. J. Geodyn. 71, 74e85. Cooper, M.J., Beevers, M.D., Oppenheimer, M., 2008. The potential impact of sea level rise on the coastal region of New Jersey, USA. Clim. Change 90, 475e492. Crichton, D., 1999. The risk triangle. In: Ingleton, J. (Ed.), Natural Disaster Management. Tudor Rose, London. CUAS, 2009. Carta dell'Utilizzazione Agricola del Suolo e Regione Campania. www. agricoltura.regione.campania.it (accessed 07.03.14.). Cutter, S.L., 1996. Vulnerability to environmental hazards. Prog. Hum. Geog 20, 529e539. Davidson-Arnott, R.G.D., 2005. A conceptual model of the effects of sea level rise on sandy coasts. J. Coast. Res. 21, 1166e1172. Di Paola, G., Iglesias, J., Rodríguez, G., Benassai, G., Aucelli, P.P.C., Pappone, G., 2011. Estimating coastal vulnerability in a meso-tidal beach by means of quantitative and semi-quantitative methodologies. J. Coast. Res. SI61, 303e308. http:// dx.doi.org/10.2112/SI61-001.30. Di Paola, G., Aucelli, P.P.C., Benassai, G., Rodríguez, G., 2014. Coastal vulnerability to wave storms of Sele littoral plain (southern Italy). Nat. Hazards 71, 1795e1819. http://dx.doi.org/10.1007/s11069-013-0980-8. EUROSION, 2004. Living with Coastal Erosion in Europe: Sediment and Space for Sustainability. www.eurosion.org (accessed 07.03.14.). FLOODsite, 2004. Evaluating Flood Risk Analysis and Management Methodologies, Language of Risk. Report: T32-04-01. Gabrielsen, P., Bosch, P., 2003. Internal Working Paper Environmental Indicators: Typology and Use in Reporting. European Environment Agency, Copenhagen. Gornitz, V.M., Daniels, R.C., White, T.M., Birdwell, K.R., 1994. The development of a coastal risk database for the U.S. Southeast: erosion and inundation form sea level rise. J. Coast. Res. (SI12), 327e338. Gornitz, V.M., Beaty, T.W., Daniels, R.C., 1997. A Coastal Hazards Data Base for the U.S. West Coast. ORNL/CDIAC-81, NDP-043C. Oak ridge national laboratory, Oak ridge, Tennessee. Grinsted, A., Moore, J.C., Jevrejeva, S., 2010. Reconstructing sea level from paleo and projected temperatures 200 to 2100 AD. Clim. Dynam. 34, 461e472. Hansen, J.E., Sato, M., 2012. Paleoclimate implications for human-made climate   change. In: Berger, A., Mesinger, F., Sija cki, D. (Eds.), Climate Change: Inferences from Paleoclimate and Regional Aspects. Springer, pp. 21e48. http://dx.doi.org/ 10.1007/978-3-7091-0973-1_2.

G. Benassai et al. / Ocean & Coastal Management 104 (2015) 22e35 IIM, 2002. Tidal Data Base. Istituto Idrografico della Marina, Genova. IPCC, 2014. Fifth Assessment Report (AR5) e Climate Change 2014: Mitigation of Climate Change e Intergovernmental Panel on Climate Change. www.ipcc.ch (accessed 07.05.14.). ISO/IEC, 2009. ISO Guide 73 e Risk Management e Vocabulary. http://www.iso.org (accessed 07.03.14.). ISPRA, 2009. La definizione del rischi costiero. http://www.isprambiente.gov.it/ contentfiles/00002900/2970-chirico.zip/at_download/file (accessed 07.03.14.). ISPRA, 2013. Data Analysis. http://www.idromare.it/analisi_dati.php (accessed 10.09.13.). ISTAT, 2001. Basi territoriali e variabili censuarie: censimenti 2001. 15eCampania. http://www.istat.it/it/files/2013/11/R15_01_WGS84.zip (accessed 07.03.14.). ISTAT, 2011. Basi territoriali e variabili censuarie: censimento 2011. 15eCampania. http://www.istat.it/it/files/2013/11/R15_11_WGS84.zip (accessed 07.03.14.). Jones, R., Boer, R., 2003. Assessing Current Climate Risks. United Nations Development Programmed, New York. http://www.undp.org/cc/pdf/APF/TP%20final/ Tech.Paper_4.qxp_30Aug04.pdf (accessed 07.08.14.). Karim, M.F., Mimura, N., 2008. Impacts of climate change and sea level rise on cyclonic storm surge floods in Bangladesh. Gloab. Environ. Change 18, 490e500. Kriebel, D.L., Dean, R.G., 1993. Convolution method for time-dependent beachprofile response. J. Waterw. Port. Coast. Ocean. Eng 119, 204e207. Kristensen, P., 2004. The DPSIR Framework. National Environmental Research Institute, Department of Policy Analysis, Denmark. Lambeck, K., Antonioli, F., Anzidei, M., Ferranti, L., Leoni, G., Scicchitano, G., Silenzi, S., 2011. Sea level change along the Italian coast during the Holocene and projections for the future. Quatern. Int. 232, 250e257. Luo, S., Wang, H., Cai, F., 2013. An integrated risk assessment of coastal erosion based on fuzzy set theory along Fujian coast, southeast Chine. Ocean Coast. Manage. 84, 68e76. http://dx.doi.org/10.1016/j.ocecoaman.2013.07.007. Martinelli, L., Zanuttigh, B., Corbau, C., 2010. Assessment of coastal flooding hazard along the Emilia Romagna littoral. Coast. Eng 57, 1042e1058. http://dx.doi.org/ 10.1016/j.coastaleng.2010.06.007. Mc Granaham, G., Bolk, D., Anderson, D., 2007. The rising tide: assessing the risk of climate change and human settlements in low elevation coastal zones. Environ. Urban 19, 17e37. Medatlas Group, 2004. Wind and Wave Atlas of the Mediterranean Sea. Western European Union, WEAO Research Cell. Nicholls, R.J., de la Vega-Leinert, A.C., 2000. Synthesis and Upscaling of Sea-leve Rise Vulnerability Assessment Studies (SURVAS): SURVAS Methodology. Flood Hazard Research Centre e Middlesex University. Nicholls, R.J., Hoozemans, F.M.J., Marchand, M., 1999. Increasing flood risk and wetland losses due to global sea-level rise: regional and global analyses. Glob. Environ. Change 9, 69e87. Overpeck, J.T., Otto-Bliesner, B.L., Miller, G.H., Muhs, D.R., Alley, R.B., Kiehl, J.T., 2006. Paleoclimatic evidence for future ice-sheet instability and rapid sea-level rise. Science 311, 1747e1750. € Ozyurt, G., Ergin, A., 2010. Improving coastal vulnerability assessments to sea-level rise: a new indicatorbased methodology for decision makers. J. Coast. Res. 26, 265e273. http://dx.doi.org/10.2112/08-1055.1. Pappone, G., Alberico, I., Amato, V., Aucelli, P.P.C., Di Paola, G., 2011. Recent evolution and the present-day conditions of the Campanian Coastal plains (South Italy): the case history of the Sele River Coastal plain. WIT Trans. Ecol. Environ. 149, 15e27. http://dx.doi.org/10.2495/CP110021.

35

Pappone, G., Aucelli, P.P.C., Alberico, I., Amato, V., Antonioli, F., Cesarano, M., Di Paola, G., Pelosi, N., 2012. Relative sea-level rise and marine erosion and inundation in the Sele river coastal plain (Southern Italy): scenarios for the next century. Rend. Lincei-Sci. Fis. 23, 121e129. http://dx.doi.org/10.1007/s12210012-0166-4. Pfeffer, W., Harper, J., O'Neel, S., 2008. Kinematic constraints on glacier contributions to 21st-century sea- level rise. Science 321, 1340e1343. http://dx.doi.org/ 10.1126/science.1159099. Pompe, J.J., Rinchart, J.R., 2008. Mitigating damage costs from hurricane strikes along the southern-eastern US Coast: a role for insurance markets. Ocean Coast. Manage 51, 782e788. Pranzini, E., 2001. Updrift river mouth migration on cuspate deltas: two examples from the coast of Tuscany (Italy). Geomorphology 38, 125e132. Purvis, M.J., Bates, P.D., Hayes, C.M., 2008. A probabilistic methodology to estimate future coastal flood risk due to sea level rise. Coast. Eng 55, 1062e1073. Rahmstorf, S., 2007. A semi-empirical approach to projecting future sea-level rise. Science 315, 368e370. Ramieri, E., Hartley, A., Barbanti, A., Duarte Santos, F., Gomes, A., Hilden, M., Laihonen, P., Marinova, N., Santini, M., 2011. Methods for Assessing Coastal Vulnerability to Climate Change. ETC-CCA, Technical Paper 1/2011. http://cca. eionet.europa.eu/docs/TP_1-2011 (accessed 07.03.14.). Rogers, S.I., Greenaway, B., 2005. A UK perspective on the development of marine ecosystem indicators. Mar. Pollut. Bull. 50 (1), 9e19. RON, 2012. Rete Ondametrica Nazionale (accessed 07.05.12.). http://www.idromare. it/analisi_ati.php. Rosskopf, C.M., Scorpio, V., 2013. Geomorphologic map of the Biferno River valley floor system (Molise, Southern Italy). J. Maps 9, 106e114. http://dx.doi.org/ 10.1080/17530350.2012.755385. Sarpkaya, T., Isaacson, M., 1981. Mechanics of Wave Forces on Offshore Structures (New York). Stockdon, H.F., Holman, R.A., Howd, P.A., Sallenger, A.H., 2006. Empirical parametrization of setup, swash, and run-up. Coast. Eng 53, 573e588. Teatini, P., Tosi, L., Strozzi, T., Carbognin, L., Cecconi, G., Rosselli, R., Libardo, S., 2012. Resolving land subsidence within the Venice Lagoon by persistent scatterer SAR interferometry. Phys. Chem. Earth 40-41, 72e79. Thieler, E.R., Hammar-Klose, E.S., 1999. National Assessment of Coastal Vulnerability to Sea-level Rise. U.S. Geological Survey, U.S. Atlantic Coast. Open-File Report, 99e593. Thywissen, K., 2006. Core terminology of disaster reduction: a comparative glossary. In: Birkmann, J. (Ed.), Measuring Vulnerability to Natural Hazardsdtowards Disaster Resilient Societies. UNUPress, Tokyo, New York, Paris. Vallefuoco, M., Lirer, F., Ferraro, L., Pelosi, N., Capotondi, L., Sprovieri, M., Incarbona, A., 2012. Climatic variability and anthropogenic signatures in the Gulf of Salerno (southern-eastern Tyrrhenian Sea) during the last half millennium. Rend. Lincei-Sci. Fis. 23, 13e23. http://dx.doi.org/10.1007/s12210-0110154-0. nches-Reales, J.M., Trottini, M., Chao, B.F., 2011. Mediterranean Sea level Vigo, M.I., Sa variation: analysis of the satellite altimetric data, 1992-2008. J. Geodyn. 52, 271e278. http://dx.doi.org/10.1016/j.jog.2011.02.002.
, S., 2009. Ground deforVilardo, G., Ventura, G., Terranova, C., Matano, F., Nardo mation due to tectonic, hydrothermal, gravity, hydrogeological, and anthropic processes in the Campania Region (Southern Italy) from Permanent Scatterers Synthetic Aperture Radar Interferometry. Remote Sens. Environ. 113, 197e212.